Diffusion Model
A type of generative artificial intelligence that creates novel outputs by learning to iteratively remove random noise from a completely unstructured signal.
Think of It Like This
Like a sculptor starting with a random block of marble and slowly chiseling away the rough edges until a perfectly detailed statue is revealed.
Diffusion models have become the state-of-the-art approach for image generation, largely surpassing Generative Adversarial Networks (GANs). They work by defining a forward process that adds noise and a reverse process that learns to denoise via a neural network. These architectures are highly stable during training and can be easily conditioned.